$0 to $2.4M in 12 months
$0 to $2.4M in 12 months. A beauty brand we cannot name, and a creative and offer system built to scale profitably and win Q4.
7-figure DTC beauty brand, under NDA
- 1OfferBundles and a bonus gift. AOV from $20 to $73.
- 2AngleHonest product application, shot native on a phone.
- 3TestingAbout five fresh creatives a day, 70% scale, 30% test.
- 4Q4$1.65M at 3.3x ROAS, $840K in the Black Friday month.
A brand-new Shopify store with zero ad and pixel data, a $20 AOV and a CPA around $33 in a crowded, competitive market. The job: the creative and offer system that could scale profitably and be ready to win Q4.
HYLN's part: the creative strategy, the testing system and the offer input. The brand's own team ran media buying, email and the site.
Offer, audience and message aligned before spending hard, then a fast launch: 25 video creatives in week one, to gather real data instead of guessing.
A better offer, built with the brand. Once we saw how the product was actually used, we pushed for quantity discounts, free shipping from two units and a free bonus gift at three or four. Ordering more became the obvious choice: AOV went from $20 to $73, up 266%, and conversion rate from 3.09% to 4.34%.
The winning angle, honest product application: native, POV, iPhone-shot demos centred on the product and the creator, opened by a strong three-second hook, with voiceover, subtitles and a clear call to action. We tested about five fresh creatives a day, 70% of budget to scale and 30% to test.
Q4. Budgets started early, in mid-October, so the winners were proven before the season. Q4 closed at $1.65M at 3.3x ROAS, with $840K in the Black Friday month.
The creative did not win on its own. The offer moved AOV before the ads scaled, and the ads had something worth scaling because of it. The lesson we carry into every account: align the offer, the audience and the message before spending hard, then test at a pace the data can answer. Five fresh creatives a day only works when the offer underneath them already converts.
